Core particle polymerization defect detection method and equipment

Abstract

The invention relates to the technical field of wafer detection, and provides a core particle polymerization defect detection method and equipment, which are used for improving the accuracy of core particle polymerization defect detection. According to the method, a unique detection mode of the core particle aggregation defect is designed for preliminary detection of normal candidate core particles, area analysis of a core particle region is performed by using a binary image of the candidate core particles under a target light source, target core particles possibly having the core particle aggregation defect are obtained, and the core particle aggregation defect is generally caused by core particle dislocation, so that the core particle aggregation defect can be accurately detected. Therefore, after the core grain boundary crossing condition is obtained according to the detection frame of the target core grain in the first detection result, the upper, lower, left and right boundary interested regions in the binary image of the target core grain are adjusted, so that the region where the misplaced core grain is located is accurately obtained, and the misplaced core grain area is determined according to the boundary core grain area proportion of the boundary interested regions. Whether the target core particle has the core particle polymerization defect or not is determined, the precision of chip polymerization defect detection is effectively improved, the omission ratio of defective products is reduced, and the quality of electronic products is further ensured.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of wafer detection, and provides a core particle aggregation defect detection method and device. BACKGROUND

[0002] At present, with the explosive growth of artificial intelligence and technology products, the quality of integrated circuits and semiconductor core particles, as core basic elements, directly determines the performance and reliability of the entire electronic product. With the development of electronic products towards miniaturization and high performance, the core particle manufacturing process is continuously improved, the core particle size is continuously reduced, and the integration level is higher and higher.

[0003] Generally, the size of a single wafer is about 6-8 inches, and the LED core particles contained therein are about tens of thousands to hundreds of thousands, and the maximum can reach one million. These core particles are regularly arranged on the wafer. When the wafer structure is distorted, especially under the action of some external environment or stress, the core particles may be misaligned, and multiple core particles may be aggregated (such as adhesion, close proximity or overlap, etc.), forming a "doublet" structure. This structure sometimes interferes with normal electron flow or signal conduction, affects the performance of the core particles, and further affects the running stability of the electronic product.

[0004] Therefore, core particle aggregation defect detection has become a problem to be solved in the development process of electronic products at present. SUMMARY

[0005] The application embodiment provides a core particle defect detection method and device, which is used for improving the accuracy of core particle defect detection.

[0006] In a first aspect, the application embodiment provides a core particle defect detection method, comprising:

[0007] obtaining a first detection result of each of a plurality of core particles contained in a wafer, and taking a core particle with an initial normal first detection result as a candidate core particle;

[0008] According to the binary image of the candidate core particle under the target light source, the core particle area of the core particle region is calculated, and when the core particle area meets a preset abnormal condition, the candidate core particle is taken as a target core particle;

[0009] According to the detection frame of the target core particle in the first detection result, a core particle boundary crossing situation is obtained, and four boundary interest regions in the binary image of the target core particle are adjusted according to the core particle boundary crossing situation; the boundary interest region is preset according to the core particle model;

[0010] When the boundary core particle area of the adjusted boundary interest region meets a preset area ratio, the first detection result of the target core particle is modified as a core particle aggregation defect.

[0011] The beneficial effects of the above technical solutions are: after the preliminary detection of the plurality of core particles contained in the wafer, the core particles with the first detection result of initial normality are taken as candidate core particles for subsequent fine detection, and the core particles with defects are filtered, thereby improving the efficiency of fine detection. Moreover, for the fine detection process of the subsequent candidate core particles, a unique detection method of core particle aggregation defects is designed. Considering that the area of the core particle region is large when the core particles are aggregated, the area analysis of the core particle region is performed by using the binary image of the candidate core particle under the target light source to obtain the target core particle that may have core particle aggregation defects. Generally, the core particle aggregation defects are caused by mispositioning of the core particles, and therefore, the four boundary regions of interest in the binary image that may have mispositioning core particles are generally preset according to the core particle model. In this way, after the core particle out-of-bound condition is obtained according to the detection frame of the target core particle in the first detection result, the four boundary regions of interest in the binary image of the target core particle are adjusted according to the core particle out-of-bound condition, so as to accurately obtain the region where the mispositioning core particles are located, and then the boundary core particle area ratio of the boundary region of interest is used to determine whether the target core particle has core particle aggregation defects, thereby effectively improving the precision of the chip aggregation defect detection, reducing the missed detection rate of defective products, and further ensuring the quality of electronic products.

[0012] Optionally, the obtaining of the core particle out-of-bound condition according to the detection frame of the target core particle in the first detection result and the adjustment of the four boundary regions of interest in the binary image of the target core particle according to the core particle out-of-bound condition comprise:

[0013] comparing the center point coordinates of the detection frame of the target core particle in the first detection result with the image origin coordinates of the target core particle; wherein the image origin is the top left corner vertex of the image;

[0014] determining the core particle out-of-bound direction according to the comparison result, and reducing the four boundary regions of interest in the binary image of the target core particle according to the core particle out-of-bound direction.

[0015] The beneficial effects of the above technical solutions are: the position of the detection frame is usually affected when the chips are aggregated, and the image origin is generally the top left corner vertex. In this way, by comparing the center point coordinates of the detection frame with the image origin coordinates, the chip out-of-bound direction can be determined, the boundary region of interest where the mispositioning core particles may be located is adjusted, and the accuracy of the core particle aggregation defect detection is ensured.

[0016] Optionally, the determining of the core particle out-of-bound direction according to the comparison result and the reducing of the four boundary regions of interest in the binary image of the target core particle according to the core particle out-of-bound direction comprise:

[0017] When the horizontal coordinate of the center point is less than the horizontal coordinate of the image origin, it is determined that the out-of-bound direction of the core particle is the left boundary of the image, the boundary region of interest of the left boundary of the image is deleted, the boundary region of interest of the right boundary of the image is kept unchanged, and the width of the boundary region of interest of the upper boundary and the lower boundary of the image is reduced;

[0018] When the horizontal coordinate of the center point is greater than the horizontal coordinate of the image origin, it is determined that the out-of-bound direction of the core particle is the right boundary of the image, the boundary region of interest of the right boundary of the image is deleted, the boundary region of interest of the left boundary of the image is kept unchanged, and the width of the boundary region of interest of the upper boundary and the lower boundary of the image is reduced;

[0019] When the vertical coordinate of the center point is less than the vertical coordinate of the image origin, it is determined that the out-of-bound direction of the core particle is the upper boundary of the image, the boundary region of interest of the upper boundary of the image is deleted, the boundary region of interest of the lower boundary of the image is kept unchanged, and the height of the boundary region of interest of the left boundary and the right boundary of the image is reduced;

[0020] When the vertical coordinate of the center point is greater than the vertical coordinate of the image origin, it is determined that the out-of-bound direction of the core particle is the lower boundary of the image, the boundary region of interest of the lower boundary of the image is deleted, the boundary region of interest of the upper boundary of the image is kept unchanged, and the height of the boundary region of interest of the left boundary and the right boundary of the image is reduced.

[0021] The technical scheme has the beneficial effects that different boundary region adjustment modes are designed according to the specific out-of-bound direction of the core particle, so that the accuracy of subsequent chip aggregation defect analysis is ensured.

[0022] Optionally, after the first detection result of the target core particle is modified as the core particle aggregation defect, the method further includes:

[0023] Obtaining a second detection result of each of a plurality of core particles contained in the wafer, and fusing the first detection result and the second detection result of each core particle to obtain a target detection result;

[0024] Ordering the plurality of core particles according to a certain rule, and screening at least one core particle with a target detection result of a core particle aggregation defect, and taking the screened at least one core particle as a to-be-analyzed core particle respectively;

[0025] For each to-be-analyzed core particle, a target detection result of the to-be-analyzed core particle is obtained according to the gap between the to-be-analyzed core particle and the adjacent plurality of core particles.

[0026] The beneficial effects of the above technical solutions are: by aggregating the detection results obtained by multiple detection methods, the detection accuracy of the core particle defects is improved, and since the use of the gap between the core particles for analyzing the core particle aggregation defects is affected by the contamination defects, which may lead to inaccurate detection, the detected core particle aggregation defects are rejudged using the gap between the core particles, thereby further improving the accuracy of the core particle aggregation defect detection.

[0027] Optionally, the target detection result of the core particle to be analyzed is obtained according to the gap between the core particle to be analyzed and the adjacent core particle, and the method comprises:

[0028] For the plurality of adjacent core particles, the gap between the core particle to be analyzed and each adjacent core particle is calculated respectively;

[0029] When the minimum gap in the plurality of gaps meets the preset gap condition, the target detection result of the core particle to be analyzed is determined as a core particle aggregation defect;

[0030] When the minimum gap in the plurality of gaps does not meet the preset gap condition, the target detection result of the core particle to be analyzed is determined as a normal chip.

[0031] The beneficial effects of the above technical solutions are: when the gap between the core particles is used for rejudging the detected core particle aggregation defects, the accuracy of the core particle aggregation defect rejudgment is ensured.

[0032] Optionally, the binary image is obtained in the following manner:

[0033] From the initial core particle images of the candidate core particle under a plurality of light sources, an initial core particle image under a target light source is selected; wherein the initial core particle image under the target light source has a gray foreground and a white background;

[0034] The pixel values of the foreground and the background in the selected initial core particle image are flipped, and the gap in the flipped initial core particle image is filled to obtain a target core particle image of the candidate core particle under the target light source;

[0035] According to the gray level histogram of the target core particle image, a target segmentation threshold value of the foreground and the background is obtained; wherein the target segmentation threshold value makes the gray level difference between the foreground and the background maximum and the gray level difference within the foreground and the background minimum;

[0036] According to the segmentation threshold value, the target core particle image is binarized to obtain a binary image of the candidate core particle.

[0037] The beneficial effects of the above technical solutions are that, since the foreground and background color difference of the initial core grain image under the target light source is large, the core grain and the background can be better distinguished, the core grain contour is more prominent after pixel value flipping and gap filling, and by setting a reasonable foreground and background segmentation threshold, the boundary between the core grain and the background is further distinguished, thereby improving the accuracy of subsequent area analysis.

[0038] Optionally, the calculating the core grain area of the core grain region according to the binary image of the candidate core grain under the target light source, and regarding the candidate core grain as the target core grain when the core grain area meets a preset abnormal condition comprises:

[0039] extracting a connected region with a pixel value of 255 from the binary image of the candidate core grain under the target light source, and regarding the connected region as the core grain region;

[0040] counting the number of pixels in the core grain region to obtain the core grain area;

[0041] regarding the candidate core grain as the target core grain when the core grain area is greater than a preset multiple of a preset core grain area.

[0042] The beneficial effects of the above technical solutions are that, when there is a core grain aggregation defect, the area of the core grain is large, and therefore, whether the candidate core grain is the target core grain of multiple core grains can be determined by comparing the area of the candidate core grain with the area of a normal core grain.

[0043] Optionally, when the area of the intermediate hole in the binary image is greater than a preset area threshold, the method further comprises:

[0044] extracting a first maximum connected region with a pixel value of 255 from the binary image, and flipping the pixel value in the first maximum connected region to 0;

[0045] extracting a second maximum connected region from the flipped binary image, and setting the pixel value in the second maximum connected region to 255 to obtain a new binary image.

[0046] The beneficial effects of the above technical solutions are that, since the structures of different types of chips can be different, some chips include a light-emitting region inside, the light-emitting region is generally close in color to the background, and the area of the core grain is generally calculated according to the core grain contour, in order to be applicable to more types of core grains, therefore, the first maximum connected region with a pixel value of 255 is flipped, so that the second maximum connected region represents the core grain region of multiple types of core grains, and the algorithm consistency is ensured.

[0047] In a second aspect, an embodiment of the present application provides a core grain defect detection device, comprising a processor, a memory and a communication interface, the communication interface, the memory and the processor are connected through a bus;

[0048] The communication interface is configured to acquire an image;

[0049] The memory stores a computer program, and the processor executes the following operations according to the computer program:

[0050] Obtain a first detection result of each of a plurality of core particles contained in a wafer, and take a core particle with an initial normal first detection result as a candidate core particle;

[0051] According to the binary image of the candidate core particle under the target light source, calculate the core particle area of the core particle region, and when the core particle area meets a preset abnormal condition, take the candidate core particle as a target core particle;

[0052] According to the detection frame of the target core particle in the first detection result, obtain a core particle out-of-bound situation, and according to the core particle out-of-bound situation, adjust four boundary interest regions in the binary image of the target core particle; the boundary interest region is preset according to the core particle model;

[0053] When the boundary core particle area of the adjusted boundary interest region meets a preset area ratio, modify the first detection result of the target core particle to a core particle aggregation defect.

[0054] Optionally, the processor obtains a core particle out-of-bound situation according to the detection frame of the target core particle in the first detection result, and adjusts four boundary interest regions in the binary image of the target core particle according to the core particle out-of-bound situation, and the specific operation is:

[0055] Compare the center point coordinates of the detection frame of the target core particle in the first detection result with the image origin coordinates of the target core particle; wherein the image origin is the top left corner vertex of the image;

[0056] According to the comparison result, determine the core particle out-of-bound direction, and according to the core particle out-of-bound direction, reduce the four boundary interest regions in the binary image of the target core particle.

[0057] Optionally, the processor determines the core particle out-of-bound direction according to the comparison result, and reduces the four boundary interest regions in the binary image of the target core particle according to the core particle out-of-bound direction, and the specific operation is:

[0058] When the center point horizontal coordinate is less than the image origin horizontal coordinate, determine that the core particle out-of-bound direction is the image left boundary, delete the boundary interest region of the image left boundary, keep the boundary interest region of the image right boundary unchanged, and reduce the width of the boundary interest regions of the image upper boundary and the image lower boundary;

[0059] When the horizontal coordinate of the center point is greater than the horizontal coordinate of the image origin, it is determined that the core particle out-of-bound direction is the right boundary of the image, the boundary region of interest of the right boundary of the image is deleted, the boundary region of interest of the left boundary of the image is kept unchanged, and the width of the boundary region of interest of the upper boundary and the lower boundary of the image is reduced;

[0060] When the vertical coordinate of the center point is less than the vertical coordinate of the image origin, it is determined that the core particle out-of-bound direction is the upper boundary of the image, the boundary region of interest of the upper boundary of the image is deleted, the boundary region of interest of the lower boundary of the image is kept unchanged, and the height of the boundary region of interest of the left boundary and the right boundary of the image is reduced;

[0061] When the vertical coordinate of the center point is greater than the vertical coordinate of the image origin, it is determined that the core particle out-of-bound direction is the lower boundary of the image, the boundary region of interest of the lower boundary of the image is deleted, the boundary region of interest of the upper boundary of the image is kept unchanged, and the height of the boundary region of interest of the left boundary and the right boundary of the image is reduced.

[0062] Optionally, after the processor modifies the first detection result of the target core particle as the core particle aggregation defect, the processor further performs:

[0063] Obtain the second detection result of each of a plurality of core particles contained in the wafer, and fuse the first detection result and the second detection result of each core particle to obtain a target detection result;

[0064] Orderly arrange the plurality of core particles according to a certain rule, and screen at least one core particle whose target detection result is a core particle aggregation defect, and take the screened at least one core particle as a to-be-analyzed core particle respectively;

[0065] For each to-be-analyzed core particle, the target detection result of the to-be-analyzed core particle is obtained according to the gap between the to-be-analyzed core particle and the adjacent plurality of core particles.

[0066] Optionally, the processor obtains the target detection result of the to-be-analyzed core particle according to the gap between the to-be-analyzed core particle and the adjacent core particle, and the specific operation is:

[0067] For the adjacent plurality of core particles, the gap between the to-be-analyzed core particle and each adjacent core particle is calculated respectively;

[0068] When the smallest gap in the plurality of gaps meets the preset gap condition, it is determined that the target detection result of the to-be-analyzed core particle is a core particle aggregation defect;

[0069] When the smallest gap in the plurality of gaps does not meet the preset gap condition, it is determined that the target detection result of the to-be-analyzed core particle is a normal chip.

[0070] Optionally, the processor obtains the binary image by:

[0071] selecting an initial core grain image under a target light source from the initial core grain images of the candidate core grain under multiple light sources; wherein the initial core grain image under the target light source has a foreground in gray and a background in white;

[0072] reversing pixel values of the foreground and the background in the selected initial core grain image, and filling the gap in the reversed initial core grain image to obtain a target core grain image of the candidate core grain under the target light source;

[0073] obtaining a target segmentation threshold of the foreground and the background according to a gray level histogram of the target core grain image; wherein the target segmentation threshold makes the gray level difference between the foreground and the background maximum and the gray level difference within the foreground and the background minimum;

[0074] performing a binarization processing on the target core grain image according to the segmentation threshold to obtain a binary image of the candidate core grain.

[0075] Optionally, the processor calculates a core grain area of a core grain region according to the binary image of the candidate core grain under the target light source, and takes the candidate core grain as a target core grain when the core grain area satisfies a preset abnormal condition, and the specific operation is as follows:

[0076] extracting a connected region with a pixel value of 255 from the binary image of the candidate core grain under the target light source, and taking the connected region as the core grain region;

[0077] counting the number of pixels in the core grain region to obtain the core grain area;

[0078] taking the candidate core grain as the target core grain when the core grain area is greater than a preset multiple of a preset core grain area.

[0079] Optionally, when the area of the intermediate hole in the binary image is greater than a preset area threshold, the processor further performs:

[0080] extracting a first maximum connected region with a pixel value of 255 from the binary image, and reversing the pixel value in the first maximum connected region to 0;

[0081] extracting a second maximum connected region from the reversed binary image, and setting the pixel value in the second maximum connected region to 255 to obtain a new binary image.

[0082] In a third aspect, an embodiment of the present application provides a computer readable storage medium, which stores computer executable instructions, and the computer executable instructions can implement the steps of the core grain aggregation defect detection method when executed.

[0083] The technical effects brought by any one of the implementation manners of the third aspect to the third aspect can refer to the technical effects brought by the corresponding implementation manners of the first aspect and the second aspect, which will not be described here. BRIEF DESCRIPTION OF DRAWINGS

[0084] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0085] Figure 1 A flowchart of a core particle polymerization defect detection method provided by the embodiments of the present application is shown in the figure.

[0086] Figure 2 A flowchart of a core particle preliminary detection method provided by the embodiments of the present application is shown in the figure.

[0087] Figure 3A A local wafer image provided by the embodiments of the present application is shown in the figure.

[0088] Figure 3B A first detection result of a core particle provided by the embodiments of the present application is shown in the figure.

[0089] Figure 4 A binarization process of a core particle image provided by the embodiments of the present application is shown in the figure.

[0090] Figure 5A And Figure 5B A foreground and background of an initial core particle image of two different types of core particles is shown in the figure.

[0091] Figure 6 An adjustment flowchart of a boundary region of interest provided by the embodiments of the present application is shown in the figure.

[0092] Figure 7 An adjustment effect of a boundary region of interest provided by the embodiments of the present application is shown in the figure.

[0093] Figure 8 A re-trial flowchart of a core particle polymerization defect provided by the embodiments of the present application is shown in the figure.

[0094] Figure 9 A core particle gap provided by the embodiments of the present application is shown in the figure.

[0095] Figure 10A A core particle image acquisition result provided by the embodiments of the present application is shown in the figure.

[0096] Figure 10B A core particle image detection result provided by the embodiments of the present application is shown in the figure.

[0097] Figure 11 A core particle polymerization defect detection process schematic diagram provided for an embodiment of the present application;

[0098] Figure 12 A structure diagram of a core particle defect detection device provided for an embodiment of the present application. DETAILED DESCRIPTION

[0099] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions of the present application will be described below in connection with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments described in the present application document, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0100] Based on the exemplary embodiments shown in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application. In addition, although the disclosure in the present application is introduced according to one or several examples, it should be understood that each aspect of these disclosures can also constitute a complete technical solution independently.

[0101] It should be noted that the brief description of the terms in the present application is only for the convenience of understanding the subsequently described embodiments, and is not intended to limit the embodiments of the present application. Unless otherwise specified, these terms should be understood according to their ordinary and general meanings.

[0102] The terms "first", "second", and the like in the specification and claims of the present application and the above-described drawings are used to distinguish similar or similar objects or entities, and do not necessarily mean a specific order or sequence, unless otherwise indicated. It should be understood that the terms used in this way can be interchanged under appropriate circumstances, for example, those other than the order given in the illustration or description of the embodiments of the present application can be implemented.

[0103] In addition, the terms "include" and "have" and any variations thereof are intended to cover but not exclusive inclusion, for example, a product or device including a series of components does not have to be limited to those components clearly listed, but can include other components not clearly listed or inherent to these products or devices.

[0104] The term "module" used in the present application refers to any known or later developed hardware, software, firmware, artificial intelligence, fuzzy logic or combination of hardware or / and software code capable of performing functions related to the element.

[0105] Core particle aggregation: refers to the phenomenon that multiple core particles are adhered, adjacent or overlapped together. When the number of aggregated core particles is 2, it is generally called "doublet", and when the number of aggregated core particles is greater than 2, it is generally called "multicore".

[0106] For the core particle aggregation phenomenon caused by core particle mispositioning, the morphology is complex and is easily disturbed by background noise, and the general detection method (such as single light source imaging, single algorithm detection) is difficult to accurately identify, often leading to false detection or missed detection.

[0107] In view of this, the core particle aggregation defect detection method provided in the embodiments of the present application designs a unique detection method for core particle aggregation defects. Considering that the area of the core particle region is large when the core particles are aggregated and the core particles may cross the boundary, the area of the core particle region is analyzed using the binary image of the initially normal candidate core particle under the target light source to obtain the target core particle that may have a core particle aggregation defect. Generally, the core particle aggregation defect is caused by core particle mispositioning, so the four boundary regions of interest in the binary image where the mispositioned core particles may exist are usually preset according to the core particle model. In this way, after the core particle boundary crossing condition is obtained according to the detection frame of the target core particle in the first detection result, the four boundary regions of interest in the binary image of the target core particle are adjusted according to the core particle boundary crossing condition, so as to accurately obtain the region where the mispositioned core particle is located, and then determine whether the target core particle has a core particle aggregation defect according to the core particle area ratio of the boundary region, thereby effectively improving the precision of the chip aggregation defect detection, reducing the missed detection rate of defective products, and further ensuring the quality of electronic products.

[0108] On the other hand, the core particles on the wafer are preliminarily detected when the core particle aggregation defect is detected, the core particles with the first detection result being initially normal are taken as candidate core particles for subsequent fine detection, and the defective core particles are filtered, thereby improving the efficiency of fine detection. The whole coarse detection and fine detection process realizes the cooperation of the artificial intelligence (AI) algorithm and the traditional visual image processing algorithm, and effectively improves the accuracy of the core particle aggregation defect detection.

[0109] Referring to Figure 1 A core particle aggregation defect detection method is provided in the embodiments of the present application, which mainly includes the following steps:

[0110] S101: Obtain the first detection result of each of the plurality of core particles contained in the wafer, and take the core particle with the first detection result being initially normal as a candidate core particle.

[0111] In some embodiments, considering that a single light source cannot highlight different types of defects at the same time, such as doublet gap and background dirt, when the core particles are preliminarily detected, the core particle images under multiple light sources are collected.

[0112] As shown in Figure 2 , it is a schematic diagram of the preliminary detection process of the core particle, mainly including the following steps:

[0113] S1011: Collect local wafer images from multiple perspectives under different light sources.

[0114] Wherein, the local wafer images from multiple perspectives under each light source can be spliced into a complete wafer image, and each local wafer image contains multiple complete core particles.

[0115] As shown in Figure 3A , it is a schematic diagram of a local wafer image, and each gray area represents a core particle.

[0116] S1012: Cut out multiple initial core particle images of each core particle from each local wafer image.

[0117] Wherein, one core particle exists in multiple initial core particle images under multiple light sources.

[0118] S1013: For each core particle, splice the initial core particle images of the core particle under multiple light sources to obtain a core particle splicing image.

[0119] Wherein, the initial core particle image under each light source occupies one channel.

[0120] S1014: Input the core particle splicing images of multiple core particles into the AI model respectively to obtain the first detection result of each core particle.

[0121] Wherein, the AI model includes but is not limited to Transform model, BERT model, YOLO model and ResNet model, etc.

[0122] In one embodiment, the first detection result contains not only the type of core particle, but also the detection frame of the core particle, the position of the chip, etc.

[0123] As shown in Figure 3B , it is a schematic diagram of the first detection result, wherein the abnormal and normal chips are circled with detection frames of different colors, and the numbers in them represent the positions of the core particles output by the AI model, that is, the first core particle in the image.

[0124] In some embodiments, the number of core particles on a wafer is large, in order to improve the detection efficiency of the system, the collected local wafer images can be uploaded to the GPU, and the core particle segmentation, detection and other processes can be performed on the GPU.

[0125] After the preliminary detection of the plurality of core particles contained in the wafer, for the abnormal core particles whose defect types have been detected, subsequent precision detection is not required. In order to avoid missing detection of core particle aggregation defects, the normal core particles after preliminary detection can be subjected to precision detection, and the core particles with defects are filtered out, thereby effectively improving the detection efficiency of the core particle aggregation defects.

[0126] S102: Calculate the core particle area of the core particle region according to the binary image of the candidate core particle under the target light source.

[0127] During core particle defect detection, one light source cannot detect all defects. Different light sources can better present different defects. Therefore, the wafer is photographed under multiple light sources. Therefore, for each core particle, a target light source that makes the core particle structure clearest can be selected from multiple light sources, and the initial core particle image corresponding to the target light source is used for defect precision detection.

[0128] In an embodiment, the core particle area is generally large when the core particles are aggregated. Therefore, the core particle aggregation defects can be identified by area analysis method on the binary image.

[0129] As shown in Figure 4 , it is a schematic diagram of the obtaining process of the binary image, mainly including the following steps:

[0130] S1021: Select an initial core particle image under a target light source from the initial core particle images of the candidate core particle under multiple light sources.

[0131] In the initial core particle image under the target light source, the foreground (i.e. the core particle) is gray and the background is white, that is, the color difference between the foreground and the background is large, so that the structure of the core particle can be clearly reflected.

[0132] In some embodiments, the structures of different types of core particles are different, so the foreground structures in the images are different.

[0133] As shown in Figure 5A and Figure 5B , they are schematic diagrams of the foreground and background of the initial core particle images of two different types of core particles, wherein, Figure 5A The core particle in contains a light-emitting area, so the foreground contains a hole consistent with the background color.

[0134] S1022: Flip the pixel values of the foreground and background in the selected initial core particle image.

[0135] For example, when the pixel value is 0, it is converted to 255, when the pixel value is 1, it is converted to 254, when the pixel value is 2, it is converted to 253, and so on.

[0136] By pixel value flipping, the seed particles (i.e. foreground) in the initial seed particle image can be shown as bright color, and the background can be shown as dark color, which facilitates subsequent image denoising and morphological operation, etc., and makes the seed particle contour more complete.

[0137] S1023: Fill the gap in the flipped initial seed particle image to obtain a target seed particle image of the candidate seed particle under the target light source.

[0138] Generally, there is noise during image acquisition, and the pixel boundary gray level is large after amplification, which affects the detection result. Therefore, the flipped initial seed particle can be smoothed to make the gray levels between pixels more smooth.

[0139] Since there is a gap in the polymerized seed particle, which affects the detection effect, for the smoothed initial seed particle image, the gap in the middle can be filled by morphological operations such as dilation and erosion, thereby improving the clarity of the seed particle contour.

[0140] S1024: Obtain a target segmentation threshold of the foreground and the background according to a gray level histogram of the target seed particle image.

[0141] The target segmentation threshold makes the gray level difference between the foreground and the background maximum, and the gray level difference within the foreground and the background minimum.

[0142] In specific implementation, all pixel points in the target seed particle image are traversed, the frequency of each gray level is counted, a gray level histogram of the target seed particle image is generated, then an initial threshold is given, the target seed particle image is divided into a class less than or equal to the initial threshold and a class greater than the initial threshold, and the inter-class variance is calculated. The initial threshold is adjusted to make the inter-class variance maximum, and the target segmentation threshold of the foreground and the background is obtained.

[0143] S1025: Perform binaryzation processing on the target seed particle image according to the segmentation threshold to obtain a binary image of the candidate seed particle. Wherein, the pixel value greater than the segmentation threshold is set to 255, and the pixel value less than or equal to the segmentation threshold is set to 0. The target seed particle image is converted into a binary image with pixel values of 0 and 255 through the segmentation threshold.

[0144] Since the foreground and the background color of the initial seed particle image under the target light source have large difference, the seed particle and the background can be better distinguished. After pixel value flipping and gap filling, the seed particle contour is more prominent. Moreover, by setting a reasonable foreground and background segmentation threshold, the boundary between the seed particle and the background is further clarified, thereby improving the accuracy of subsequent area analysis.

[0145] In some embodiments, for the seed particle with a high-light region at the center of the image, as shown in FIG. 8, the middle region is a large hole after binaryzation, therefore, the hole filling can be performed on the binary image of such seed particle. Figure 5A In some embodiments, for the seed particle with a high-light region at the center of the image, as shown in FIG. 8, the middle region is a large hole after binaryzation, therefore, the hole filling can be performed on the binary image of such seed particle.

[0146] In specific implementation, when the area of the intermediate hole in the binary image is greater than the preset area threshold, a first maximum connected region with pixel value of 255 in the binary image is extracted, the pixel value in the first maximum connected region is flipped to 0, a second maximum connected region in the flipped binary image is extracted again, the pixel value in the second maximum connected region is set to 255, and a new binary image is obtained.

[0147] Because the structures of different types of chips may be different, some chips contain a light-emitting region inside, and the light-emitting region is generally close in color to the background. The chip area is generally calculated according to the chip contour. In order to be suitable for more chip types, the first maximum connected region with pixel value of 255 is flipped, so that the second maximum connected region represents the chip region of multiple chip types, thereby ensuring the consistency of the algorithm.

[0148] The boundary between the chip and the background in the binary image is relatively clear, so the chip area can be calculated according to the chip region in the binary image. In specific implementation, a connected region with pixel value of 255 is extracted from the binary image of the candidate chip under the target light source, and the connected region is taken as the chip region. The number of pixels in the chip region is counted to obtain the chip area.

[0149] S103: When the chip area meets the preset abnormal condition, the candidate chip is taken as the target chip.

[0150] When there is a chip aggregation defect, the area of the chip is larger, so whether the candidate chip is the target chip of multiple chips can be determined by comparing the area of the normal chip. In specific implementation, the width and height of the normal chip can be input in advance to obtain the normal chip region, and the number of pixels in the normal chip region is extracted to obtain the preset chip area. When the extracted chip area in the binary image is greater than a preset multiple (such as 1.15) of the preset chip area, it indicates that the candidate chip is abnormal, so the candidate chip is taken as the target chip, otherwise it indicates that the candidate chip is a normal chip.

[0151] S104: According to the detection frame of the target chip in the first detection result, the chip out-of-bound situation is obtained, and four boundary interest regions in the binary image of the target chip are adjusted according to the chip out-of-bound situation.

[0152] When multiple chips aggregate, there may be out-of-bound chips due to misplacement of the chips, which may cause the chip image to contain a half chip, so the boundary interest region where the chip may appear can be set in advance at the upper, lower, left and right four boundaries of the binary image according to the chip type, and the four boundary interest regions are adjusted according to the actual chip out-of-bound situation.

[0153] Generally, the size of a single core particle is small, and the aggregation of two core particles is common when the core particle is cut. For example, in a "doublet" structure, the adjustment process of the boundary interest region, as shown in Figure 6 includes the following steps:

[0154] S1041: Compare the center point coordinates of the detection frame in the first detection result with the image origin coordinates of the target core particle.

[0155] The image origin is the top-left vertex of the image.

[0156] S1042: Determine the core particle out-of-boundary direction according to the comparison result.

[0157] Specifically, when the horizontal coordinate of the center point is less than the horizontal coordinate of the image origin, the core particle out-of-boundary direction is determined to be the left boundary of the image; when the horizontal coordinate of the center point is greater than the horizontal coordinate of the image origin, the core particle out-of-boundary direction is determined to be the right boundary of the image; when the vertical coordinate of the center point is less than the vertical coordinate of the image origin, the core particle out-of-boundary direction is determined to be the upper boundary of the image; and when the vertical coordinate of the center point is greater than the vertical coordinate of the image origin, the core particle out-of-boundary direction is determined to be the lower boundary of the image.

[0158] S1043: Reduce the four boundary interest regions in the binary image of the target core particle according to the core particle out-of-boundary direction.

[0159] As shown in Figure 7 , the boundary interest region is filled with diagonal lines. When the core particle out-of-boundary direction is the left boundary of the image, the boundary interest region of the left boundary of the image is deleted, the boundary interest region of the right boundary of the image is kept unchanged, and the width of the boundary interest region of the upper boundary and the lower boundary of the image is reduced; when the core particle out-of-boundary direction is the right boundary of the image, the boundary interest region of the right boundary of the image is deleted, the boundary interest region of the left boundary of the image is kept unchanged, and the width of the boundary interest region of the upper boundary and the lower boundary of the image is reduced; when the core particle out-of-boundary direction is the upper boundary of the image, the boundary interest region of the upper boundary of the image is deleted, the boundary interest region of the lower boundary of the image is kept unchanged, and the height of the boundary interest region of the left boundary and the right boundary of the image is reduced; and when the core particle out-of-boundary direction is the lower boundary of the image, the boundary interest region of the lower boundary of the image is deleted, the boundary interest region of the upper boundary of the image is kept unchanged, and the height of the boundary interest region of the left boundary and the right boundary of the image is reduced.

[0160] For example, the original size of the boundary interest region is 50 pixels wide and 50 pixels high. For left-right crossing, the width of the boundary interest region of the upper and lower boundaries is changed to 30 pixels. For up-down crossing, the height of the boundary interest region of the left and right boundaries is changed to 30 pixels.

[0161] The position of the detection frame is usually affected when the chips are aggregated, and the image origin is generally the top-left corner. Therefore, by comparing the center point coordinates of the detection frame with the image origin coordinates, the out-of-bound direction of the chip can be determined, and different boundary region adjustment methods can be designed according to the specific out-of-bound direction of the chiplet, so as to accurately obtain the position of the misaligned chiplet and ensure the accuracy of the chip aggregation defect detection.

[0162] S105: When the area of the adjusted boundary region of interest meets the preset area ratio, the first detection result of the target chiplet is modified as a chip aggregation defect.

[0163] For each remaining boundary region of interest after adjustment, a gray scale analysis is performed to determine the boundary chiplet area in the boundary region of interest. When the boundary chiplet area in the remaining boundary region of interest accounts for more than 50% of the corresponding boundary region of interest, it indicates that the target chiplet does not have a "doublet" structure, and therefore the first detection result of the target chiplet is modified as a chip aggregation defect. Otherwise, the result of the target chiplet as a normal chiplet remains unchanged.

[0164] The embodiments of the present application design a unique detection method for chip aggregation defects. Since the area of the chip region is large when the chips are aggregated, the area of the chip region is analyzed using the binary image of the candidate chiplet under the target light source to obtain the target chiplet that may have a chip aggregation defect. Chip aggregation defects are generally caused by misaligned chiplets. Therefore, the boundary regions of interest where misaligned chiplets may appear are set at the upper, lower, left and right boundaries according to the chip model, and the four boundary regions of interest are adjusted according to the out-of-bound condition of the chiplet determined by the detection frame in the first detection result, so as to accurately obtain the region where the misaligned chiplet is located. Then, according to the boundary chiplet area ratio of the boundary region of interest, it is determined whether the target chiplet has a chip aggregation defect, which effectively improves the precision of chip aggregation defect detection, reduces the miss rate of defective products, and further ensures the quality of electronic products.

[0165] In some embodiments, the background of some chiplets is accompanied by large-area dirt, as shown in FIG. 1B, which may affect the detection result of chip aggregation. In order to further improve the accuracy of chip aggregation defect detection, the chip aggregation defect can be rejudged. Figure 3B

[0166] As shown in FIG. 1C, the rejudgment process of the chip aggregation defect mainly includes the following steps: Figure 8

[0167] S106: Obtain the second detection result of each chiplet contained in the wafer, and fuse the first detection result and the second detection result of each chiplet to obtain a target detection result.

[0168] ​​Generally, the core particle defect detection includes appearance defect detection and internal defect detection. The appearance defect detection is generally obtained through a visual image, that is, the first detection result in the embodiment of the application, and the internal defect is generally obtained through electrical measurement (for example, measurement through a pen), which is recorded as a second detection result. Finally, the target detection result of the core particle (that is, a defective core particle or a normal core particle) needs to be fused with the detection results obtained in multiple ways. This process can be referred to as file merging. Through aggregation of the detection results obtained in multiple ways, the detection accuracy of the core particle defect is improved.

[0169] S107: orderly arranging the plurality of core particles according to a certain rule.

[0170] The core particles are arranged in a certain rule during wafer production, and the gap between the core particles is fixed. However, during preliminary detection, the shooting angle and the detection frame output by the AI model are shaken, which causes each core particle in the pixel space not to be arranged in a neat manner, as shown in FIG. 7A. Therefore, the plurality of core particles need to be sorted in the local wafer and global wafer space. Each core particle after sorting has its own position coordinates, and the adjacent core particles can also be obtained. Figure 3B

[0171] S108: screening at least one core particle with a target detection result of core particle aggregation defect, and taking the screened at least one core particle as a core particle to be analyzed.

[0172] In order to avoid over-detection of the core particle aggregation defect, the core particle with the core particle aggregation defect can be taken as the core particle to be analyzed for rejudgment.

[0173] S109: for each core particle to be analyzed, obtaining a target detection result of the core particle to be analyzed according to the gap between the core particle to be analyzed and the plurality of adjacent core particles.

[0174] The core particle aggregation defect is caused by misplacement of the core particle. After the core particle is misplaced, the gap between the adjacent core particles changes. Therefore, for each core particle to be analyzed, after obtaining the core particles adjacent to the upper, lower, left and right of the core particle to be analyzed from the plurality of sorted core particles, whether the core particle to be analyzed really has the core particle aggregation defect can be determined through the gap between the core particles.

[0175] Taking one core particle to be analyzed as an example, the gap between the core particle to be analyzed and the core particles adjacent to the upper, lower, left and right is calculated, as shown in FIG. 8A. One minimum gap is selected from the four gaps, and when the minimum gap meets a preset gap condition, it is determined that the target detection result of the core particle to be analyzed is the core particle aggregation defect. When the minimum gap does not meet the preset gap condition, it is determined that the target detection result of the core particle to be analyzed is a normal chip. Figure 9

[0176] ​​The minimum gap can be either the longitudinal gap between the upper and lower core particles or the transverse gap between the left and right core particles. Each of the longitudinal and transverse gaps corresponds to a gap threshold. When the minimum gap is a longitudinal gap, the preset gap condition is that the minimum gap is 1 / 2 of the normal upper and lower core particle gap. When the minimum gap is a transverse gap, the preset gap condition is that the minimum gap is 1 / 2 of the normal left and right core particle gap.

[0177] It should be noted that the embodiments of this application do not impose restrictive requirements on the gap threshold in the preset gap conditions, and can be flexibly adjusted according to the actual situation.

[0178] When analyzing core particle polymerization defects by core particle area, the influence of contamination defects may lead to inaccurate detection. Therefore, for the core particles to be analyzed that have detected core particle polymerization defects, the gap between the core particles is used to re-judge the core particle polymerization defects, thereby further improving the accuracy of core particle polymerization defect detection.

[0179] like Figure 10A The image shows initial images of multiple core particles. The detection results for normal and abnormal core particles can be marked in different ways (e.g., using different colored detection boxes). Abnormal core particles with different defect types (e.g., dirt, core particle aggregation, scratches, short circuits, etc.) can also be distinguished in different ways. Figure 10B As shown, the embodiments of this application can accurately detect core particle polymerization defects.

[0180] See Figure 11 The diagram below shows the detection architecture for core polymerization defects provided in this application embodiment. The main process is as follows:

[0181] S1: Acquire local wafer images.

[0182] Because of the small size of the die, a single light source may not be able to reveal the defects of the die. Therefore, a high-definition industrial camera can be used to acquire local wafer images from different perspectives under multiple light sources.

[0183] S2: Perform target detection on the local wafer image to obtain the detection box for each chip.

[0184] S3: The wafer image is cropped according to the detection frame to obtain the chip image.

[0185] S4: For each core image, AI detection algorithm and CV detection algorithm are used to detect defects in the core image respectively.

[0186] The AI ​​detection algorithm includes logic for detecting core particle aggregation defects.

[0187] S5: For each core image, the AI ​​detection results and CV detection results are fused.

[0188] S6: Sort the detected core grain image by rows and columns.

[0189] S7: Rejudge the core grain with core grain aggregation defects.

[0190] S8: Output the final detection result according to the rejudgment result.

[0191] When the rejudgment is a core grain aggregation defect, the detection result does not need to be modified and can be directly outputted. When the rejudgment is not a core grain aggregation defect, the detection result is modified and then outputted.

[0192] Those skilled in the art can understand that various aspects of the present application can be implemented as a system, a method or a program product. Therefore, various aspects of the present application can be embodied as a complete hardware embodiment, a complete software embodiment (including firmware, microcode, etc.), or an embodiment combining hardware and software aspects, which can be collectively referred to as "circuitry", "module" or "system" here.

[0193] Based on the same technical concept, the present application provides a core grain defect detection device, which can realize the steps of the core grain aggregation defect detection method and achieve the same technical effect.

[0194] Referring to Figure 12 The gateway device includes a processor 1201, a memory 1202 and a communication interface 1203, the communication interface 1203, the memory 1202 and the processor 1201 are connected through a bus 1204;

[0195] The communication interface 1203 is configured to acquire an image.

[0196] The memory 1202 stores a computer program, and the processor 1201 executes Figure 1 the steps of the core grain aggregation defect detection method according to the computer program.

[0197] It should be noted that Figure 12 is only an example, and the core grain defect detection device not shown can also include memory, power supply, crystal oscillator, camera and other devices.

[0198] In the embodiments of the present application, the memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system and programs required for running instant messaging functions, etc.; and the data storage area can store various instant messaging information and operation instruction sets, etc. The memory can be a volatile memory such as a random-access memory (RAM), or a non-volatile memory such as a read-only memory, a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD), or any other medium capable of carrying or storing desired computer programs in the form of instructions or data structures and capable of being accessed by a computer, but is not limited thereto. The memory can be a combination of the above memories. The processor can include one or more central processing units (CPUs), GPUs or digital processing units, etc.

[0199] The specific connection medium between the communication interface, the memory and the processor is not limited in the embodiments of the present application. In the embodiments of the present application, the bus between the communication interface, the memory and the processor is described by a thick line in the figure, and the connection mode between other components is only schematically described and is not limited thereto. The bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of description, only one thick line is described in the figure, but only one bus or only one type of bus is not described.

[0200] The embodiments of the present application also provide a computer-readable storage medium for storing some instructions, which can complete the steps of the core particle polymerization defect detection method in any of the foregoing embodiments when executed.

[0201] The embodiments of the present application also provide a computer program product for storing a computer program, which is used for executing the steps of the core particle polymerization defect detection method in any of the foregoing embodiments.

[0202] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can adopt a computer program product in the form of being implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program codes.

[0203] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0204] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0205] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0206] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A core pellet polymerization defect detection method characterized by, The method comprises: obtaining first detection results of each of a plurality of core particles contained in a wafer, taking core particles with the first detection results being initial normal as candidate core particles; calculating a core particle area of a core particle region according to a binary image of the candidate core particle under a target light source, and taking the candidate core particle as a target core particle when the core particle area meets a preset abnormal condition; obtaining a core particle out-of-bound condition according to a detection frame of the target core particle in the first detection result, and adjusting four boundary interest regions in the binary image of the target core particle according to the core particle out-of-bound condition; the boundary interest regions are preset according to a core particle model; when a boundary core particle area of the adjusted boundary interest regions meets a preset area proportion, modifying the first detection result of the target core particle to a core particle aggregation defect.

2. The method of claim 1, wherein, The method of obtaining a core particle out-of-bound condition according to a detection frame of the target core particle in the first detection result, and adjusting four boundary interest regions in the binary image of the target core particle according to the core particle out-of-bound condition comprises: comparing a center point coordinate of the detection frame of the target core particle in the first detection result with an image origin coordinate of the target core particle; the image origin is a top left corner vertex of the image; determining a core particle out-of-bound direction according to the comparison result, and reducing the four boundary interest regions in the binary image of the target core particle according to the core particle out-of-bound direction.

3. The method of claim 2, wherein, The method of determining a core particle out-of-bound direction according to the comparison result, and reducing the four boundary interest regions in the binary image of the target core particle according to the core particle out-of-bound direction comprises: when the horizontal coordinate of the center point is less than the horizontal coordinate of the image origin, determining that the core particle out-of-bound direction is an image left boundary, deleting the boundary interest region of the image left boundary, keeping the boundary interest region of the image right boundary unchanged, and reducing the width of the boundary interest regions of the image upper boundary and the image lower boundary; when the horizontal coordinate of the center point is greater than the horizontal coordinate of the image origin, determining that the core particle out-of-bound direction is an image right boundary, deleting the boundary interest region of the image right boundary, keeping the boundary interest region of the image left boundary unchanged, and reducing the width of the boundary interest regions of the image upper boundary and the image lower boundary; when the vertical coordinate of the center point is less than the vertical coordinate of the image origin, determining that the core particle out-of-bound direction is an image upper boundary, deleting the boundary interest region of the image upper boundary, keeping the boundary interest region of the image lower boundary unchanged, and reducing the height of the boundary interest regions of the image left boundary and the image right boundary; when the vertical coordinate of the center point is greater than the vertical coordinate of the image origin, determining that the core particle out-of-bound direction is an image lower boundary, deleting the boundary interest region of the image lower boundary, keeping the boundary interest region of the image upper boundary unchanged, and reducing the height of the boundary interest regions of the image left boundary and the image right boundary.

4. The method of claim 1, wherein, After the first detection result of the target core particle is modified to a core particle aggregation defect, the method further comprises: obtaining second detection results of each of a plurality of core particles contained in the wafer, and fusing the first detection result and the second detection result of each core particle to obtain a target detection result; The plurality of core particles are arranged in order according to a certain rule, and at least one core particle with a target detection result of core particle aggregation defect is screened out, and the at least one screened core particle is taken as a to-be-analyzed core particle respectively; For each to-be-analyzed core particle, the target detection result of the to-be-analyzed core particle is obtained according to the gap between the to-be-analyzed core particle and the adjacent plurality of core particles.

5. The method of claim 4, wherein, The target detection result of the to-be-analyzed core particle is obtained according to the gap between the to-be-analyzed core particle and the adjacent core particle, comprising: For the plurality of adjacent core particles, the gap between the to-be-analyzed core particle and each adjacent core particle is calculated respectively; When the smallest gap in the plurality of gaps meets the preset gap condition, it is determined that the target detection result of the to-be-analyzed core particle is a core particle aggregation defect; When the smallest gap in the plurality of gaps does not meet the preset gap condition, it is determined that the target detection result of the to-be-analyzed core particle is a normal chip.

6. The method of any one of claims 1-5, wherein, The binary image is obtained in the following manner: From the initial core particle image of the candidate core particle under a plurality of light sources, an initial core particle image under a target light source is selected; wherein the initial core particle image under the target light source has a gray foreground and a white background; Flip the pixel values of the foreground and background in the selected initial core particle image, and fill the gap in the flipped initial core particle image to obtain a target core particle image of the candidate core particle under the target light source; According to the gray level histogram of the target core particle image, a target segmentation threshold value of the foreground and the background is obtained; wherein the target segmentation threshold value makes the gray difference between the foreground and the background maximum, and the gray difference within the foreground and the background minimum; According to the segmentation threshold value, the target core particle image is binarized to obtain a binary image of the candidate core particle.

7. The method of claim 6, wherein, The core particle area of the core particle region is calculated according to the binary image of the candidate core particle under the target light source, and when the core particle area meets the preset abnormal condition, the candidate core particle is taken as a target core particle, comprising: From the binary image of the candidate core particle under the target light source, a connected region with a pixel value of 255 is extracted, and the connected region is taken as a core particle region; The number of pixels in the core particle region is counted to obtain a core particle area; When the core particle area is greater than a preset multiple of a preset core particle area, the candidate core particle is taken as a target core particle.

8. The method of claim 6, wherein, When the area of the intermediate hole in the binary image is greater than a preset area threshold, the method further comprises: Extracting a first maximum connected region with a pixel value of 255 in the binary image, and flipping the pixel value in the first maximum connected region to 0; Extracting a second maximum connected region in the flipped binary image, setting the pixel value in the second maximum connected region to 255 to obtain a new binary image.

9. A core particle defect detection apparatus characterized by comprising: It comprises a processor, a memory and a communication interface, the communication interface, the memory and the processor are connected through a bus; The communication interface is used to acquire an image; The memory stores a computer program, and the processor executes the following operations according to the computer program: Obtain the first detection result of each of the plurality of core particles contained in the wafer, and take the core particle with the initial normal first detection result as a candidate core particle; According to a binary image of the candidate core particle under a target light source, a core particle area of a core particle region is calculated, and the candidate core particle is taken as a target core particle when the core particle area meets a preset abnormal condition; According to a detection frame of the target core particle in the first detection result, a core particle border crossing situation is obtained, and four boundary interest regions in a binary image of the target core particle are adjusted according to the core particle border crossing situation; the boundary interest regions are preset according to a core particle model; When a boundary core particle area of the adjusted boundary interest regions meets a preset area proportion, the first detection result of the target core particle is modified as a core particle aggregation defect.

10. The core pellet defect inspection apparatus of claim 9, wherein After the processor modifies the first detection result of the target core particle as the core particle aggregation defect, the processor further performs: obtaining second detection results of a plurality of core particles contained in the wafer respectively, and fusing the first detection result and the second detection result of each core particle to obtain a target detection result; sequentially arranging the plurality of core particles according to a certain rule, and screening at least one core particle whose target detection result is the core particle aggregation defect, and taking the screened at least one core particle as a to-be-analyzed core particle respectively; for each to-be-analyzed core particle, obtaining a target detection result of the to-be-analyzed core particle according to a gap between the to-be-analyzed core particle and a plurality of adjacent core particles.

Images